There are four types of measures that should be tracked if you want to improve medication safety.

Process measures. These measures help assess how well you are performing core processes associated with medication use. Measuring core processes helps determine if there is variation in carrying them out, which could lead to undesirable outcomes, and if there are preventable risks associated with processes, which could result in harm. Process measures can be identified for all facets of medication use. However, high-volume and high-risk processes, or processes associated with high-alert medications, should be targeted to maximize the benefit to patient safety. A few examples include:

One newly evolving measure to evaluate improvement within a process is to track the total risk priority number (RPN) of a process that has undergone a failure mode and effects analysis. As an organization works to improve the process, the RPN should decrease over time as effective changes are implemented. For more information on this process measure, visit www.ihi.org (click on Topics, Patient Safety, Medication Systems, Measures).

Structure measures. These measures assess the organizational culture, values, and leadership. They differ from process measures in that they are not task oriented, but foundational in nature. Examples include:

The Agency for Healthcare Research and Quality (AHRQ) has recently released a new, tested survey for assessing the safety culture in hospitals. The tool also offers guidance on sampling, data collection, and analysis of findings (see www.ahrq.gov/qual/hospculture/ for details).

Outcome measures. These measures assess whether your efforts to improve medication safety have been successful. As such, many believe that medication errors are the most useful outcome measure for medication safety. However, harm is a much more reliable and powerful measure, especially if it keeps you intellectually engaged with the possibility that all harm is preventable.

If errors are used to measure medication safety, self-reporting is the typical data-gathering tool, which is highly inaccurate. Errors are the obvious focus, so any adverse drug events uncovered are quickly sorted into preventable and “non-preventable” categories. This, in turn, promotes the tacit acceptance of “non-preventable” harm as a property of the medication system, something for which you have no responsibility. On the other hand, if harm is used as a measure of medication safety, the measure is reliable, clear, and direct, and the focus is on all unintended results. This keeps you intellectually engaged with the possibility of reducing all patient harm, to admit that you can do better, to raise the bar when it comes to patient safety. For example, most hospitals collect data on readmissions to the hospital. If bleeding episodes from warfarin caused some of the readmissions, these events may not be fully assessed if the focus is on errors alone. In these cases, errors may not be apparent, so the events would likely be tagged as “non-preventable” adverse drug reactions. But if focused only on preventing harm in patients who take warfarin, not error, you’re more likely to explore ways to reduce all occurrences of bleeding. Thus, the best outcome measure for medication safety is all adverse drug events, regardless of causation.

Evaluating patient records using a list of triggers is the most effective means of collecting data on adverse drug events. Triggers are clues that an adverse drug event may have occurred. Follow up is needed for confirmation. An extensive list of triggers can be found here.

Examples include:

Computerized methods for detecting adverse drug events via a trigger tool have proved very effective (Classen DC, Metzger J. Improving medication safety: the measurement conundrum and where to start. Internat Jour for Qual in Health Care 2003; 15:i41-i47). However, this methodology requires a high level of automation with customized software linkage to clinical databases, so the initial outlay may be costly. However, recently, a relatively low cost, “low tech” method for using a trigger tool to uncover adverse drug events has been devised and tested in more than 80 hospitals (Rozich JD, Haraden CR, Resar RK. Adverse drug event trigger tool: a practical methodology for measuring medication related harm. Qual Saf Health Care 2003; 12;194-200). The technique, which requires minimal training, appears to increase the rate of adverse drug event detection 50-fold over traditional reporting methodologies.

Balancing measures. These measures are used to ensure that a change in one part of the system is not causing problems in another part of the system. For example, by using balancing measures, one hospital quickly learned that a change in antiemetics to reduce the time a patient must spend in the oncology clinic actually resulted in reduced patient satisfaction. Patients felt rushed and unable to talk to staff about their diagnosis and therapy.

Measuring medication safety is not easy, but it must be a core component of your improvement efforts. If you do not have an effective measurement plan in place, an interdisciplinary team should consider the examples above and identify a place to start. Be sure to clearly describe each measure, its goal, and the data collection plan. Remember: traditional efforts to measure medication safety have not been successful in guiding improvement. Thus, even if you currently have a measurement plan in place, it may be time to look at it again with fresh eyes and updated tools.